背　景
許多醫院手術室開刀作業排程，是以排程人員（多為手術室護理長）的經驗以及資源限制來決定，往往造成超時的工作或資源利用缺乏效率。

目　的
提出一個綜合量化與質化多目標考量之手術室開刀作業排程模型，促進排程的品質。

方　法
本研究應用模糊德爾菲法求得時間、成本、與風險因子的權重；利用模糊理論將風險因子量化；利用渴望函數，整合時間、成本、與風險因子，發展手術室作業排程綜合績效評量模型，並使用模擬退火演算法建構排程系統，驗證模型之績效。

結　果
綜合績效評量模型成功的將量化與質化指標整合為單一量化指標。驗證結果顯示，應用綜合績效評量模型結合模擬退火演算法，可得到最有效率的手術室作業排程，並有助於控制成本與降低疲勞風險。

結　論／實務應用
將成本、疲勞風險等重要因素納入手術室作業排程考量，可使所得排程更客觀且績效更佳。應用本研究成果時，應將手術室每間房間的成本可能並不相同列入考量；亦建議可增加其它因子（如加班成本、護理人員疲勞風險），使排程績效評量模型更接近實際狀況。本研究所使用的模型建構程序與方法，可作為未來實務上發展手術室作業排程系統之參考。

Background: Most operating room (OR) scheduling is done by scheduling personnel based on experience or heuristics. Such often results in excessive overtime for medical personnel and poor resource-use efficiencies.

Purpose: This study was designed to employ an integrated OR scheduling performance evaluation model (IOSPEM), which integrates quantitative and qualitative multiple objectives,to the improvement of OR scheduling quality.

Methods: The fuzzy Delphi method was used to obtain the weight values of time factor, cost factor, and risk factor. The fuzzy set theory was applied to transform qualitative risk measures into quantitative values. The desirability function was utilized to integrate time, cost, and risk factors to develop the IOSPEM. The simulated annealing algorithm was used to develop the scheduling system and test proposed model performance.

Results: The proposed IOSPEM successfully integrated the quantitative and qualitative indices into a sole quantitative index. Experiment results show that the IOSPEM incorporating the simulated annealing algorithm is able to obtain the most efficient OR schedule and is helpful in reducing costs and fatigue risks.

Conclusions and Implications for Practice: Operating room scheduling will be made more objective and efficient if OR scheduling personnel can simultaneously consider the cost,fatigue risk, and other factors in scheduling. Cost of each OR room should be considered to set appropriate cost coefficients in practical application of the IOSPEM. It is also suggested that other indices (e.g., OR overtime costs and OR nurse fatigue risks) also be considered in the proposed model so as to better reflect the actual scheduling environment. The procedure and methods implemented in this study may be used as the basis for further developing more effective and efficient OR scheduling systems.